Effectiveness, acceptability and usefulness of mobile applications for cardiovascular

disease self-management: systematic review with meta-synthesis of quantitative and

qualitative data.

Genevieve M. Coorey,a,b Lis Neubeck,c John Mulley,a Julie Redfern,a,b

a The George Institute for Global Health, Australia

b Sydney Medical School, University of Sydney, Sydney, Australia

c School of Health and Social Care, Edinburgh Napier University, Edinburgh, UK

Funding: This research received no specific grant from any funding agency in the public,

commercial, or not-for-profit sectors. GMC is supported by a National Heart Foundation

Health Professional Scholarship (101544). JR is funded by a Career Development and Future

Leader Fellowship co-funded by the National Health and Medical Research Council and the

National Heart Foundation (APP1061793).

Competing interests: None.

Ethical approval: Not required.

Corresponding author and requests for reprints:

Genevieve Coorey

The George Institute for Global Health, Australia

PO Box M201 Missenden Rd, Camperdown

Sydney, NSW 2050 Australia

Ph. +61 2 8052 4644

Fax: +61 2 8052 4501

gcoorey@georgeinstitute.org.au

Word count (including references): 7662

1

ABSTRACT

Background: Mobile technologies are innovative, scalable approaches to reducing risk of

cardiovascular disease (CVD) but evidence related to effectiveness and acceptability remains

limited. We aimed to explore the effectiveness, acceptability and usefulness of mobile

applications (apps) for CVD self-management and risk factor control.

Design: Systematic review with meta-synthesis of quantitative and qualitative data.

Methods: Comprehensive search of multiple databases (Medline, Embase, CINAHL,

SCOPUS, and Cochrane CENTRAL) and grey literature. Studies were included if the

intervention was primarily an app aimed at improving at least two lifestyle behaviours in

adults with CVD. Meta-synthesis of quantitative and qualitative data was performed to

review and evaluate findings.

Results: Ten studies of varying designs including 607 patients from 5 countries were

included. Interventions targeted hypertension, heart failure, stroke and cardiac rehabilitation

populations. Factors that improved among app users were rehospitalisation rates, disease-

specific knowledge, quality of life, psychosocial well-being, blood pressure, body mass

index, waist circumference, cholesterol, and exercise capacity. Improved physical activity,

medication adherence, and smoking cessation were also characteristic of app users.

Appealing app features included tracking healthy behaviours, self-monitoring, , disease

education, and personalised, customisable content. Small samples, short duration, and

selection bias were noted limitations across some studies, as was the relatively low overall

scientific quality of evidence.

Conclusions: Multiple behaviours and CVD risk factors appear modifiable in the shorter

term with use of mobile apps. Evidence for effectiveness requires larger, controlled studies

of longer duration, with emphasis on process evaluation data to better understand important

system- and patient-level characteristics.

2

PROSPERO Registration Number: CRD42017068482

Abstract word count: 249

KEYWORDS: cardiovascular disease, mobile applications, mHealth, lifestyle behaviour,

systematic review

3

INTRODUCTION

Cardiovascular disease (CVD) is the leading cause of death worldwide, accounting for

approximately one third of deaths.1 This burden is borne unequally, with most deaths

occurring in low- and middle-income countries2 despite a trend in higher income countries

towards lower age-related CVD mortality over recent decades.1 As the major contributor to

the broader burden of non-communicable disease prevalence, reducing the incidence,

morbidity and mortality of CVD is a key global health priority.3 In parallel, the World Heart

Federation has initiated a ‘25 by 25’ campaign to reduce premature mortality from CVD by

25% by 2025.4 Central to these imperatives are innovative, cost-effective and scalable

approaches that reduce behavioural and metabolic influences on CVD risk, for example

physical inactivity, smoking, overuse of sodium, and medication non-adherence.1 Long term

success with such lifestyle-related decision-making is multifactorial. In modern society, this

likely includes the array of personal mobile-based technologies aiming to support patients

living with CVD.5-8

The prevalence of mobile technology offers an obvious delivery medium for expanding the

reach of secondary prevention. Smartphone ownership among adults was reported in 2015 at

60% and 72% in Europe and the United States, respectively,9 far exceeding the global median

of 45%. Worth noting is that the usage rate of 37% across emerging/developing economies,9

for example in sub-Saharan Africa, South Asia and parts of Central America, lessens the

reach of this potentially beneficial technology. In 2016, mobile health app downloads were

estimated at 3.2 billion.10 Recent figures suggest that mobile devices (smartphones and

tablets) account for 49% of Internet traffic,10 underscoring their convenience and practicality.

Mobile technologies are increasingly recognised for other benefits, such as bridging time and

distance barriers to clinical oversight, and expanding accessibility to care that is traditionally

4

delivered face-to face, thereby reducing evidence-practice gaps.11, 12 Recent systematic

reviews have studied the effectiveness of multi-component technology-based interventions in

CVD prevention. Typically these comprise two or more of: interactive web sites, email, tele-

monitoring, video, one- or two-way text messaging, phone-based applications (apps), and

non-phone devices for data acquisition prior to wireless upload to a central monitoring hub.13-

15 Other reviews have targeted a single delivery format, for example web sites16 or text

messaging17; or have targeted a population with a specific diagnosis, for example heart

failure18 or coronary heart disease.19 Multi-component interventions often incur greater

resource needs, especially if linked into a centralised clinical hub. On the other hand, text-

messaging alone under-utilises the range of interactive capabilities within modern mobile

devices. The interactive traits that patients feel are essential may not only be those that

facilitate contact with health professionals. At present, there is a lack of qualitative research

data about which mobile app features and functions are engaging and interesting over time

for the goal of changing multiple behaviours.13 Moreover, determining patient groups for

whom the content or other components require modification in order to be more effective

requires further investigation.19

Overall, more research about mobile apps is needed to address the gaps emerging from recent

reviews in which mobile-based apps are under-represented. Therefore, the aim of this study

was to review the potential effectiveness, acceptability and usefulness of patient-directed

applications for CVD self- management that are delivered and used primarily on a

smartphone or tablet device. Of interest is: (1) whether chiefly standalone apps with low

reliance on day-to-day clinician involvement can improve risk factor control and disease self-

care; and (2) patients’ perspectives on preferred app features and perceived utility for

supporting treatment adherence and healthier lifestyle behaviour.

5

METHODS

Study Design

Systematic review with meta-synthesis of quantitative and qualitative data. The review was

conducted in accordance with the Preferred Reporting Items for Systematic Reviews and

Meta-Analyses (PRISMA) guidelines.20

Search strategy

A comprehensive database search for randomised and non-randomised studies to 29th April

2017 was carried out in the Cochrane Central Registry of Controlled Trials (CENTRAL),

Medline, Embase, CINAHL and SCOPUS. Relevant studies were also sought from two trial

registers: www.clinicaltrials.gov and www.anzctr.org.au (Australian New Zealand Clinical

Trials Registry). Reference lists of other reviews and meta-analyses were manually searched

to find additional studies. The following terms were searched: mobile communication, mobile

applications, cell phone, mobile health, mHealth, mobile app, internet app, web app, smart

phone, iPhone, android, cardiovascular disease. The full search strategy is provided as a

supplementary file. No language or date restrictions were applied.. Where possible, authors

were contacted to obtain full text publications of relevant abstracts; studies published only as

an abstract were excluded.

Study inclusion criteria

Eligible studies were those which assessed an intervention comprising a patient-directed app

primarily delivered and used on a smartphone or tablet device, with the aim of: (1) improving

adoption and/or maintenance of at least two lifestyle behaviours, such as medication

6

adherence, increased physical activity, or smoking cessation; (2) improving treatment

adherence and disease self-care behaviour, for example through use of interactive self-

monitoring and information resources. The focus was mobile applications as either a stand-

alone intervention or as the central component for the intervention. Studies were excluded if

the intervention was one- or two-way short message service only; video conferencing; real-

time, remote monitoring (telehealth); telephone calls between patient and health professional;

a web site alone; or recording and uploading measurements alone.

The target population was adults aged eighteen years or older with diagnosed CVD requiring

pharmacologic and non-pharmacologic treatment and risk factor reduction to prevent disease

recurrence or complications (termed secondary prevention).21. Primary prevention

populations were excluded. Both quantitative and qualitative data were sought for evaluation

and synthesis; therefore, randomised, non-randomised and qualitative-only study designs

were accepted. No restrictions on sample size or follow-up duration were applied; however,

studies were excluded if the app intervention was not used by participants in their home

setting.

Quantitative outcomes of interest included the following measures, where available at

baseline and last reported follow up: blood pressure (BP), weight, body mass index (BMI),

exercise capacity, physical activity (PA) level, dietary improvement, smoking cessation,

medication adherence, participation in cardiac rehabilitation, change in quality of life (QoL),

glycated haemoglobin (HbA1c), total and low-density lipoprotein (LDL) cholesterol,

rehospitalisation, cardiovascular event rates, and app usage metrics. Qualitative evaluation

outcomes of interest were patient perspectives on utility, preferences, benefits and drawbacks

of app interventions.

7

Study selection process

Titles and/or abstracts of all studies identified using the search strategy and additional sources

were screened by one reviewer (GMC) to identify those that potentially met the inclusion

criteria. The full texts of relevant studies were obtained and independently assessed for

inclusion or exclusion by two reviewers (GMC and JM). Disagreements about eligibility of

studies were resolved through discussion between reviewers or by a third reviewer (JR) until

consensus was reached.

Assessment of study quality

The quality appraisal of included studies was undertaken using one of three instruments,

depending on study design. GMC assessed each study; LN appraised 30% of studies to

ensure assessments were consistent. Using the Cochrane Risk of Bias Tool,22 the main

sources of systematic bias in randomised trials were assessed: selection bias, performance

bias, detection bias, attrition bias and reporting bias. Before and after studies were assessed

(but not excluded) using the nine-item Joanna Briggs Institute Critical Appraisal Checklist for

Quasi-Experimental Studies (non-randomised experimental studies).23 Notably, five of the ten

domains in this tool concern a comparison group, which was absent in three of the four

included studies of this type. Qualitative/observational studies were assessed using the

Critical Appraisal Skills Programme (CASP) tool.24

Data extraction, synthesis and analysis

Meta-analysis was not possible due to the variation in study designs and/or insufficient data

within studies. Therefore, a narrative synthesis was undertaken after the results of each

individual study were analysed and summarised. A customised data extraction form was used

8

to extract data for assessment of study quality and evidence synthesis. One author (GMC)

extracted the following details from each included study: location and setting, methodology,

participant characteristics, description of the intervention and comparator, duration of follow-

up, outcome data, and methodological quality. A second author (LN) verified a random

selection of extracted details. Quantitative findings were categorised as follows into nine

outcome areas from the range of findings reported in the studies: hospital readmissions, QoL,

psychosocial wellbeing, CVD risk factors, medication adherence, cardiac rehabilitation

uptake, adherence and motivation, and process evaluation measures. Not all studies reported

data for all outcomes.

RESULTS

Study selection and characteristics

A total of 1354 records were screened for possible inclusion and 101 full-text manuscripts

were reviewed for eligibility (Figure 1). All reviewed manuscripts were in English. One

abstract published in English was from a non-English journal and the full text was

unobtainable. Ten papers representing nine studies from five countries (655 recruited

participants) were eventually included (Table 1). Two studies reported different data from the

same study cohort, with one detailing quantitative results25 and the other describing a

qualitative evaluation.26 CVD populations targeted in each of the included studies were as

follows: cardiac rehabilitation or acute myocardial infarction (AMI) or acute coronary

syndromes (ACS);27-30 hypertension;25 heart failure (HF);31 stroke;32 coronary heart disease

(CHD) or HF;33 hypertension, angina or AMI.34 The 10 included studies were of different

designs: three uncontrolled before-after studies25, 32, 33; one controlled before-after study;30

three RCTs;28, 29, 31 two observational pilot studies;19, 34 and one qualitative-only study.26

Follow-up duration ranged from two weeks to six months; participant drop-out was high in

9

several studies. Mean participant age across nine studies was 59.8+6.9 years and one study34

did not report participant age. All studies reported sex of

the participants and overall 27% were female. Although

the number of participants recruited across all studies

was 655 (sample size range 10-166), baseline data

indicated there were 607 participants at this time-

point due to withdrawals prior to baseline

assessments.

10

Records identified through database searching: n=1641

Medline n=733Embase n=113

Cochrane n=154Scopus n=543CINAHL n=98

Records after 298 duplicates removed (n=1354)

Records screened (n=1354)

Scr

eeni ng

Incl

ude d

Elig

ibili ty

Ide

ntif

icat ion Additional records identified

through other sources: n=11Trial registries n=2Hand searching n=9

Records excluded: n=1253Unsuitable on title/abstract

screen n=1248No full-text available n=5

Full-text articles assessed for eligibility

(n=101)Full-text articles excluded:

n=91Incorrect intervention n=47Not target population n=20

Does not report outcomes n=24(E.g. review article, study

protocol, or letter)

Studies included (n=10)

Figure 1. Flow diagram of included studies

Study quality

Methodological quality varied within the different study designs included in this review

(supplementary file). In the three RCTs, assessed using the Cochrane Risk of Bias tool, 22

blinding of participants was impossible due to the nature of the interventions and blinding of

outcome assessors was unclear. Only one RCT described random sequence generation and

allocation concealment procedures.29 Attrition bias was assessed as high in one study29 and

low in two studies.28, 31 Reporting bias across the three trials was assessed as unclear,28 low,31

and high.29

The four non-randomised before-after studies were appraised against the Joanna Briggs

Institute Critical Appraisal Checklist for Quasi-Experimental Studies (non-randomised

experimental studies).23 Three studies25, 32, 33 had no comparison/control group so the appraisal

domains related to between-group comparisons were least well fulfilled. No studies reported

multiple pre-exposure measurements of the outcomes but two studies reported multiple post-

exposure measurements.25, 32 The domain of follow up completion, description and analysis

was most fully met by one study.25 All four studies described appropriate statistical measures.

Three qualitative studies26, 27, 34 were evaluated using the CASP instrument.24 Each one met

four of the ten quality domains. The domains of recruitment strategy and consent were

unclear in one study,34 and in all three studies the relationship between researcher and

participants was deemed uncertain. The domains of data analysis and statement of findings

were partially fulfilled across the three studies.

11

Table 1: Characteristics of included studies

Author,

year & country

CVD

population;

treatment

setting

Inclusion criteria Sample

size

Mean age

range)

%

Female

Key components of intervention

(and comparator where applicable)

Operating system and development

Study design;

duration of

follow-up

*Bengtsson et

al.

2016

Sweden

HT

Primary

health care

Medically treated

for HT; age >30;

Internet access on

mobile phone;

Swedish literacy.

50 59

(33-81)

48 Interactive Self-Management Support System: self-

reporting BP, pulse, symptoms, side effects,

medication intake and lifestyle/well-being support;

graphical data displayed on companion web site.

Timing of reminders determined by patient.

App operating system: not specified

App development: researchers, patients and

clinicians involved but details not described in this

paper; communication platform developed by

Circadian Questions 21st Century Mobile.

Uncontrolled

before-after

study.

8 weeks

Dithmer et al.

2016

MI/

angina/HT

Not specified -

stated as ‘heart

patients’.

10 48-89

(mean not

reported)

40 Tablet-based Heart Game built for Android OS;

aimed as adjunct to rehabilitation; patient and team

mate complete challenges related to healthy

Single group

observational

pilot; no before-

12

Denmark Outpatients

lifestyle; accumulate points; leader board, medals,

comparison of scores across teams.

App operating system: Android

App development: prototype built using data from

workshops, field observation and interviews with

patients and nurses

study data.

2 weeks

Forman et al.

2014

USA

CR

Outpatients

Currently enrolled

in or recently

completed phase

II CR;

User of iPhone,

iPad or iPod;

English-speaking.

26 59

(43-76)

23 Heart Coach mobile app; daily task list based on

behavioural goals and educational CR content. Text

and video educational material; prompts were both

standard (e.g. walking) and personalised (e.g.

medication); activity tracking, biometric screening

and surveys. Monitoring of task completion by CR

staff. Feedback messages were in-app and from CR

staff.

App operating system: iOS for the study; now

available on Android devices.

App development: app built on ‘Wellframe’

(Boston, MA) platform using content from the

Single group

observational

pilot; no before-

study data.

30 days

13

hospital CR protocol and then verified by CR

providers

Hägglund et al.

2015

Sweden

HF

Outpatients

(readmitted

patients

could keep

using app).

Hospitalised for

HF; discharged to

primary care; no

participation in

nurse-led HF

clinic.

Treated with

regular dose or

PRN diuretic.

72 75

(range not

reported)

32 Intervention: Home Intervention System comprised

of tablet wirelessly connected to weigh scales; also,

HF info & lifestyle advice. Activities: monitor

weight and symptoms, self-titrate diuretic dose,

complete daily tasks; view graphical displays,

visual analogue scale for evaluating perceived

health status. Clinic and tech support contact

details.

Control: HF info and clinic phone number.

App operating system: not specified

App development: not described; software

provided by Care Ligo Optilogg Systems.

Multicentre

RCT.

3 months

*Hallberg et al.

2016

HT

Primary

health care

Medically treated

for HT; age >30;

Internet access on

mobile phone;

49 60

(37-81)

47 Interactive Self-Management Support System: self-

reporting BP, pulse, symptoms, side effects,

medication intake and lifestyle/well-being support;

Companion web site offers graphed feedback of

Qualitative.

Single time-

14

Sweden Swedish literacy. data. Timing of reminders determined by patient.

App operating system: not specified

App development: researchers, patients and

clinicians involved but details not described in this

paper; communication platform developed by

Circadian Questions 21st Century Mobile.

point

15

Johnston et al.

2016

Sweden

MI

Outpatients

Diagnosis of MI;

prescribed

ticagrelor during

hospitalisation

and pre-

randomisation;

>18 years; access

to & skill with a

smartphone;

Swedish literacy;

willing to attend

follow-up visits.

166 58

(range not

reported)

19 In addition to standard CR:

Intervention: web-based smartphone app with

extended drug adherence e-diary plus secondary

prevention education modules on exercise, weight

and smoking cessation; self-entry of BP, LDL-C,

BGL data; personalised educational in-app

messages; traffic light system for feedback about

adherence status.

Control: simplified drug adherence e-diary alone.

App operating system: not specified

App development: AstraZeneca and ScientificMed

Tech.

Multicentre

RCT.

6 months

Layton et al.

2014

USA

HF/CAD

Outpatients

Inpatient but

eligible for

discharge within 3

days.

16

HF=6

CAD=10

55

(26-84)

25 Smartphone (iOS) app to supplement outpatient

discharge instructions. Daily ‘to-do’ list with

educational info, appointment & medication

reminders, PA prompts consistent with CR

program, symptom monitoring; weekly phone

survey with study personnel re rehospitalisation,

Uncontrolled

before-after

study.

60 days

16

symptoms, outpatient program use. Study team

monitors uploaded biometric data and sends 1-2

messages/week.

App operating system: iOS

App development: Wellframe Inc. Cambridge, MA;

states provider dashboard is compliant with US

Health Insurance Portability and Accountability Act

(HIPAA) medical data privacy and security

provisions.

Seo et al.

2015.

Korea

Stroke

Outpatients

Previous stroke +

>1 vascular risk

factors; Able to

use an Android

smartphone

48 52

(range not

reported)

25 Korea University Health Monitoring System for

Stroke (KUHMS) mobile app; patient enters BP,

waist circumference, BGL, smoking, exercise and

medication adherence data. Data that exceed

predefined levels trigger a message alarm to patient.

App operating system: Android

App development: Korea University

Uncontrolled

before-after

study.

6 months

Varnfield et al. MI Post-MI referred

to CR; able to

94 55 13 Intervention: Care Assessment Platform of Cardiac

Rehabilitation (CAP-CR) used instead of clinic-

Multicentre

17

2014

Australia

Outpatients

attend CR.

Experience with a

smartphone.

(range not

reported)

based CR: smartphone app with health and exercise

monitoring (step counter, BP monitor, weight), text

messages, audio and video educational and

motivational content, wellness diary (e.g., sleep,

smoking, alcohol consumption, stress); web portal

for uploaded data to be viewed by mentor prior to

weekly progress call with patient.

Control: traditional clinic-based CR

App operating system: not specified

App development: not described but multiple

vendors listed for provision of the various

monitoring apps and diary components installed on

the smartphone; content aligned with national CR

guidelines.

RCT

6 weeks

Widmer et al.

2015

ACS

Outpatients

Post-PCI.

Pre-CR or 3

months post-CR.

76 66

(range not

reported)

27 Intervention: Standard CR plus Personal Health

Assistant (PHA) online or smartphone-based app:

daily tasks based on CR guidelines for healthy

lifestyle behaviour; track progress, log weight, BP,

Controlled, non-

randomised

before-after

18

USA lab values, daily PA, diet. Interactive health status

info; educational CR info, email reminders,

graphical displays.

Comparison groups (2): standard CR alone.

App operating system: Android and iOS

App development: Mayo Clinic and Healarium, Inc.

study.

3 months

Abbreviations: CVD, cardiovascular disease; HT, hypertension; BP, blood pressure; MI, myocardial infarction; CR, cardiac rehabilitation; HF, heart failure; PRN, pro re nata; RCT, randomised controlled trial; LDL-C, low density lipoprotein cholesterol; BGL, blood glucose level; CAD, coronary artery disease; PA, physical activity; ACS, acute coronary syndromes; PCI, percutaneous coronary intervention. *Linked studies.

19

Quantitative outcomes

Primary endpoints differed between studies from user engagement and satisfaction with the

mobile application,27 the extent of cardiac rehabilitation uptake and completion,29 and

reduction in risk factors.32 As a result, studies were less amenable to direct comparison of app

effectiveness. No studies reported on new CVD event rates. Selected outcomes are

summarised in Table 2. Detailed synthesis of results for hospital readmissions, QoL and

wellbeing, CVD risk factors, medication adherence, cardiac rehabilitation uptake and

adherence, and process evaluation measures are detailed below.

Hospital readmissions

One RCT31 and one non-randomised study30 reported lower hospital readmission associated

with the intervention. Hägglund et al31 reported that hospital days per patient for heart failure

were 1.3 and 3.5 in the intervention and control groups, respectively; 62% reduction in the

intervention group (risk ratio 0.38; 95% confidence interval: 0.31-0.46, p<0.05). Heart failure

hospitalisations were 34% of all hospital days for the intervention group but 68% for the

control group. Similarly, Widmer et al30 noted significantly reduced re-hospitalisations in the

patients using the intervention concurrently with standard cardiac rehabilitation (5/25 (20%)

vs 11/19 (58%), p=0.01), and in those using it post-rehabilitation (-28%, p=0.04). In the data

review, we noted incongruence between the denominators in the table describing the study

group assignments and the authors’ data description in the text, making interpretation of their

data difficult.

20

Quality of life and psychosocial well being

QoL was assessed in four studies29-31, 35 using one of three different instruments (the EQ-5D,

the Dartmouth QoL Survey, and the SF-36). In three RCTs, results were mixed: Johnston et

al28 reported no statistically significant difference between allocation groups with AMI; the

increase in QoL scored on the EQ-5D in the intervention group was higher than for the

control group but the difference was not statistically significant. In contrast, Varnfield et al29

found that after six weeks, AMI participants in the intervention group improved their scores

on the EQ-5D significantly more than the control group (p<0.05), noting that the convenience

of home-based rehabilitation may have contributed to this result. In a controlled before-after

study30 within-group and between group Dartmouth QoL survey scores significantly

improved in ACS patients using the intervention concurrently with standard cardiac

rehabilitation, compared with controls ( p=0.009 and p=0.04, respectively). In the third

RCT,31 mental and physical scores for HF patients on the SF-36 were comparable between

intervention and control groups. However, when heart failure-specific self-care and health-

related QoL were assessed, the intervention group significantly improved scores on the

European Heart Failure Self-Care Behaviour Scale, p<0.05; and achieved significantly higher

clinical mean summary score (p<0.05) and improved physical limitation (p<0.05) on the

health-related QoL Kansas City Cardiomyopathy Questionnaire. Thus, the general QoL

improvement, where apparent, appeared to be more evident in the shorter term in CHD

patients. Disease-specific QoL improved in a HF population over three months.

One RCT29 and one controlled before-after study30 assessed psychosocial wellbeing and

found improvements in patients using the mobile apps. Varnfield et al29 reported that cardiac

rehabilitation patients in the intervention group demonstrated significantly reduced DASS-

21

Depression score (p<0.001) and DASS-Anxiety score (p=0.003) from baseline to week six,

although the between group difference was not significant. Interestingly, the K10

psychological distress score that improved in the intervention group in the shorter-term

follow-up at week six (p=0.001) remained significantly improved at month six. Widmer et

al30 reported that stress scores significantly improved in ACS patients using the intervention

concurrent with standard cardiac rehabilitation (-1.3±1.3, p=0.008).

CVD risk factors

One non-randomised study and two RCTs assessed aspects of lipid profile, with mixed

results. In a controlled before-after study30 the intervention group had significant reduction in

total cholesterol (-46.9±38.3 mg/dL, p<0.0001), LDL cholesterol (-36.7±35.7 mg/dL,

p=0.0004), and triglycerides (-39.3±69.1 mg/dL, p=0.03) compared with baseline. Patients in

the control group had a significant reduction in LDL cholesterol, however the between group

difference was not reported. Two RCTs28, 29 assessed LDL cholesterol. Meta-analysis was not

possible with the way the data were reported. Johnston et al28 reported no mean change in

LDL cholesterol between treatment groups. Varnfield et al29 reported that the imputed lipid

profile data exceeded 21% for six-week data; no six-month values are shown. The

intervention group showed no difference in LDL cholesterol at week six. It was possible to

calculate between-group differences in mean LDL-cholesterol; however, no difference was

seen.

Change in BP was assessed in three before-after studies25, 30, 32 and one RCT,29 with

improvements from baseline reported in each study. In one uncontrolled before-after study, a

statistically significant decrease was reported between the baseline mean systolic BP and

22

mean diastolic BP and the mean values at week eight (SBP, 7 mmHg; SD 18; 95% CI, 1.94-

12.25; t[48]=2.77 [p=.008]) and (DBP, 4.9 mmHg; SD 10; 95% CI, 1.95-7.8; t[48]=3.35

[p=.002]).25 In the second study, 68.8% of participants reached BP target and the change from

baseline was both significant (p=0.031) and sustained at 180 days.32 A controlled before-after

study reported systolic BP significantly improved in the patients using the intervention (-

10.8±13.5 mmHg, p=0.0009); also, systolic BP decreased significantly (p=0.01) compared

with the control group.30 One RCT found a significant between-group difference in adjusted

mean diastolic BP at week six (p=0.03); however, BP results were not reported for the later

data point at month six.29

Two studies reported change in waist circumference. A slight but significant within-group

reduction was observed in the intervention group of an RCT.29 Target waist circumference

was achieved in 77.1% of patients in an uncontrolled before-after study.32 Four studies

reported change in BMI and/or weight, with mixed results. Across two RCTs, there was no

change in BMI between groups reported by Johnston et al28 however, slight but significant

improvement in weight in the intervention group was observed by Varnfield et al29 Two

before-after studies reported significant improvement in baseline BMI.30,32

One uncontrolled before-after study32 assessed change in HbA1c, reporting that 54.2% of

participants with diabetes mellitus at baseline achieved significant improvement in this

indicator at end of study (p=0.012). Three studies that reported change in participant self-

reported smoking status28, 30, 32 found positive changes but not statistically significant results.

In one RCT, the intervention group showed more quitters than the control group.28 In an

uncontrolled before-after study32 smoker rate reduced in those who used the app longer

compared with those who used the app less often. Also, a controlled before-after study30

23

reported all smokers at baseline were non-smokers at end of study, regardless of allocation

group.

Self-reported physical activity behaviour was assessed in five studies28-30, 32, 33 and overall,

users of the applications improved on this indicator. Two RCTs reported increased PA in the

intervention participants although not statistically significant28, 29 and 89% of intervention

patients who adhered to the intervention self-entered a record of daily exercise and/or used

the step counter to reach their exercise goals.29 Two uncontrolled before-after studies had

limited data: Layton et al33 reported that patients who used the intervention up to 60 days

performed the recommended PA 42% of the time compared with 25% in those who used the

intervention for 1-30 days. Seo32 reported that at 180 days, adherence in the previous month

to moderate-intensity exercise was 10.40±9.92 days but this was not assessed at baseline to

compare the potential change. In a controlled before-after study30 the intervention group

significantly improved minutes of weekly exercise (148.1±78.5 mins/week, p<0.0001) but the

between group difference was not significant.

Two studies that focused on cardiac rehabilitation assessed change in exercise capacity. An

RCT29 reported that both study groups improved the six-minute walk test from baseline to

week six and maintained this improvement to month six. Participant dropouts reduced sample

size such that significant between-group differences were not demonstrated. The Bruce

Protocol treadmill test was used in a controlled before-after study30 in which the intervention

group significantly improved exercise capacity (2.5±2.7 ml O2/min/kg, p=0.004); between

group difference was not reported.

Medication adherence

24

Three studies25, 28, 32 assessed medication adherence by different methods and reported an

overall positive impact. In one of two uncontrolled before-after studies,25 adherence was self-

reported and responses were checked against National Prescription Repository data. The

percentage of filled prescriptions corresponded to at least 80% of the prescribed doses during

the study period. In the second study,32 patients reported days of medication adherence in the

last 30 days at 29.25, however no baseline data were reported with which to compare. In a

RCT,28 self-reported drug non-adherence at six months using the e-diary app intervention was

significantly lower in the intervention group compared with the control group (p=.025).

Cardiac rehabilitation uptake, adherence and motivation

Several studies used a cardiac rehabilitation population for the research, but only two

studies27, 29 had attitudes towards, and participation in, rehabilitation as an outcome focus.

Forman et al27 reported that users of the smartphone app in cardiac rehabilitation phase II had

a 42% lower visit cancellation rate compared with non-users. Furthermore, users in phase III

cardiac rehabilitation reported improved sense of connection to clinic staff and sustaining

goals around wellness behaviours. In a RCT,29 cardiac rehabilitation uptake was defined as

attending baseline assessment and at least one gym session (control group) or one upload of

exercise data (intervention group using app-based cardiac rehabilitation). Uptake was higher

in the intervention group: (48/60, 80% vs 37/60, 62%) (RR 1.30; 95% CI 1.03-1.64; p<0.05).

Adherence was defined as clinic cardiac rehabilitation attendance to four weeks (control

group) or uploading four weeks' exercise data (intervention group, and attending the six-week

assessment (both groups). Intervention participants were 1.4 times more likely to adhere (RR

1.4; 95% CI 1.13-1.70; p<0.05). Completion was defined as attendance at six-week

25

assessment and was 33% higher in intervention participants (RR 1.71; 95% CI 1.30-2.27;

p<0.05).

26

Table 2. Summary of selected outcomes by study design

Outcome Number of

studies that

assessed this

outcome

Open label randomized

controlled trial

Quasi-experimental studies

Controlled

before-after study

Uncontrolled

before-after study

Study Effect Study Effect Study Effect

Hospital readmissions

2Hägglund et al,

2015+++

Widmer et al,

2015+++

QoL 4

Johnston et al,

2016++

Widmer et al,

2015+++

Varnfield et al,

2014+++

Hägglund et al,

2015+++

Psychosocial well-being

2*Varnfield et al,

2014

+++ Widmer et al,

2015+++

Total cholesterol 2 Varnfield et al, x Widmer et al, +++

27

2014 2015

LDL cholesterol 3

Johnston et al,

2016-

Widmer et al,

2015+

Varnfield et al,

2014x

BP 4Varnfield et al,

2014+++

Widmer et al,

2015+++

Bengtsson et al,

2016+++

Seo et al, 2015 +++

BMI and/or weight 4

Johnston et al,

2016-

Widmer et al,

2015+++ Seo et al, 2015 +++

Varnfield et al,

2014+++

Smoking 3Johnston et al,

2016++

Widmer et al,

2015+ Seo et al, 2015 ^

Exercise capacity 2

Varnfield et al,

2014 +Widmer et al,

2015+++

Physical activity 5 Johnston et al, ++ Widmer et al, +++ Layton et al, 2014 ∫

28

2016

2015Varnfield et al,

2014∫ Seo et al, 2015 ∫

Medication adherence

3

Johnston et al,

2016 +++

Bengtsson et al,

2016∫

Seo et al, 2015 ∫

CR uptake 2Varnfield et al,

2014+++

Forman et al,

20142∫

Abbreviations: QoL, quality of life; LDL, low density lipoprotein; BP, blood pressure; BMI, body mass index; CR, cardiac rehabilitation.

Key: +++ statistically significant effect; ++ greater improvement in intervention group than control but between group difference not significant; + significant improvement in both groups but between group difference not reported or not significant; - no reported change between treatment groups; × imputed data and/or data not shown; ^ within-group improvement not significant; ∫ adherence improvement data from participant survey.

*Anxiety and psychosocial distress scores.

29

Process measures and qualitative outcomes

User engagement

Intervention uptake was defined and reported differently in each of eight studies25, 27, 28, 30-34

that described metrics such as app usage frequency, duration, data registration, or

responsiveness of the user to daily tasks. Combined with often low participant numbers, drop

outs and short exposure duration, conclusions about engagement are difficult to draw.

Completion of tasks within the app (for example, an educational module) was a typical

indicator of use in studies with an overall healthy lifestyle focus.27, 33, 34 In others, emphasis

was on logging medication intake or physical measurements.25, 28, 30-32 Forman et al19 gauged

engagement by patient completion of at least one prescribed daily task from an average of 6.3

tasks per day. Overall, patients completed on average 78% of 189 tasks set over 30 days; but

the proportion of patients completing specific tasks varied by type of task, for example PA or

disease education. Layton et al33 reported that ten of sixteen study participants withdrew prior

to day sixty but patients who completed 31-60 days used various interactive aspects of the

app overall more than those who withdrew prior to day 31. Self-reported health status on day

of hospital discharge, as well as breath sounds by physical exam, were reported correlates of

application use, with lower uptake in more medically unwell patients. The authors did not

comment on specific differences by diagnostic group of the study participants (CHD or HF).

In a short study with patients of mixed CVD diagnoses,34 eight of the 10 participants used the

game-style app challenges daily for 14 days but used the leader board feature third-daily.

In one RCT, the proportion of patients who prematurely stopped using the e-diary app was

reported as low but the actual figure was not stated.28 In another RCT,31 adherence was

defined as the number of days the patient interacted with the system, divided by the number

30

of days equipped with the system. Median adherence was 88% and suggested to be a proxy

indicator of system usability. Seo et al32 defined app usage as days with risk factor data entry

into the app. Days entering data averaged 60.4 (range 1-180). Participants were dichotomized

as compliant (entered data on >47 days) and non-compliant (entered data on <47 days);

however, achievement rates for the assessed outcomes – achieved within the first three

months and maintained over the next three months - did not differ between ‘compliant’ and

‘non-compliant’ groups. Therefore, higher frequency of data entry may not equate with

achieving improvements and other factors are likely to contribute.

In an uncontrolled before-after trial25 participants were required to log daily BP

measurements for 55 days; however, it was not clear whether all patients logged BP on each

of 55 days. Widmer et al30 assessed usage frequency by number of log-in days divided by

total number of active days and counted frequency of log-ins per week. Factors associated

with usage frequency were minutes of weekly exercise at 90 days (r2=0.24, p=0.04),

reduction in BP (r2=0.38, p=0.04) and stress scores (r2=0.32, p=0.02), and improvement in

diet score (r2=0.41, p=0.007). Thus, data entry as an indication of app engagement is a

common metric, but requires high user motivation. It may underestimate other ways patients

use app features that do not register data in a central record under research conditions.

User preferences and feedback

Five studies26-28, 33, 34 explored patient preferences (Box 1) and one study reported feedback

from cardiac rehabilitation staff.27 Patients in one small study of both HF and CHD

participants33 liked medication reminders and PA information. However, they disliked being

unable to self-enter appointments and other reminders (study team initiated these or they

31

were system-generated), and felt daily requirements for data entry or other responses were

inconvenient. In a second small qualitative study34 however, daily reminders and challenges

about required exercise and lifestyle changes were well-received but participants would have

liked puzzle-style, memory and psychological challenges alongside the standard dietary and

exercise focus.

Motivational messages were varyingly received with some patients describing them as

stimulating and encouraging, whilst others opted to not receive them. Preferred features were

to be able to formulate one’s own messages, or receive messages that more closely reflect

goal achievements or non-achievements. An example would be for a message to reinforce

benefits of having achieved their goal, or motivate them to increase their effort if required.26

Software needs would thus be more sophisticated than currently used, wherein a bank of

prepared messages is employed and in-app modification of the messages in response to user

data entry does not occur. The app used in one RCT28 did have a version of more tailored

responsiveness of the messaging system according to input from the patient. That study

reported patient satisfaction feedback but not specifically how this adaptable message logic

was rated.

Patients with hypertension found the smartphone format more convenient than a computer-

based app for self-reporting data or responding to learning tasks.26 They suggested an editing

option for self-entered data; also, to insert comments beside measurements so that any

irregular readings could be seen in the context of other things happening in their lives that

day/week. This is important for helping patients understand the relationship between lifestyle

and general health, and their BP, and thus helps them address day to day influences on BP.

32

Mixed responses were obtained to ease of reading and interpreting the graphical BP data.

Some patients relied on viewing these with nursing staff during a routine clinic visit,26 so

preferred that graphs be viewable on a smartphone, not just on an optional secure web site.

Where apps were used as an adjunct to clinic-based care, patients valued personalisation and

flexibility. Feedback from one small study identified that cardiac rehabilitation content would

ideally be adaptable to the patient’s stage of rehabilitation and overall health, rather than

identical for each patient.34

Impact on disease knowledge and participation in treatment

In a small observational study of patients in cardiac rehabilitation, 96% of participants felt

that using the app supported their participation by helping them identify personal lifestyle

challenges and goals such as healthier eating or smoking cessation.27 Further, it helped them

adhere to rehabilitation (93%) and improved the quality of clinic visits (71%). Overall, 83%

of patients had a positive experience with the app. This same study reported app feedback

from staff. Among the benefits identified were reduced cardiac rehabilitation barriers, better

patient participation, improved between-visit communication and better attendance.27

Participants in a small study using gamification design34 felt the team-based approach to

rehabilitation was effective for relatives wishing to provide emotional support. Ideally the

app should allow communication between teams, perhaps underscoring the value placed on

peer support between those with shared conditions.

33

Patients using a hypertension management app reported greater insight into their condition,

evidenced by (1) adhering to treatment even when feeling well and their BP was controlled;

(2) making lifestyle changes to positively affect BP, for example losing weight, quitting

smoking and increasing PA; and (3) being more active in discussions with their doctor.26

Limited participant feedback was reported in two of the RCTs. Johnston et al28 used a self-

reported system usability score (not shown) that was reported higher at study visit two and

end of study (p.001). Of the intervention participants, 97.5% would recommend the app to

others; 68.4% were willing to continue using the app; and 80% felt the app was relevant and

helpful for motivation. In a second RCT,29 feedback questionnaires at week six and month six

about cardiac rehabilitation exercise adherence noted that more than 85% of intervention

patients found the step counter to be motivational. More extensive feedback about using an

app to undertake the rehabilitation was not reported.

34

Box 1. Preferred app features from qualitative data in the included studies.

35

Healthy eating and exercise goal setting

Recognition of achievements

Memory and psychological tasks

Enable user editing of self-entered numeric data, reminders and

appointments

Motivational messages with:

o Opt-out option

o User-created and system-generated content

o Content responsive to user input to app

Game-based design techniques

Enable textual data to be entered with numeric data

Ensure graphical data displays are viewable on a smartphone

Tailor content of cardiac rehabilitation-related apps to stage of recovery

Offer team-based competition options

Remove requirement for daily data entry

Provide in-app “how to” guides.

DISCUSSION

This study reveals mixed findings for the effectiveness of patient-focused apps across a range

of outcomes for disease self-management and risk factor control. Factors that improved

among users of the apps were hospital readmission rates, disease-specific knowledge, general

and health-related QoL, psychosocial well-being, BP, BMI, waist circumference, cholesterol,

and exercise capacity. Improved daily PA, medication adherence (including medication self-

titration), and smoking cessation were also characteristic of app users. One uncontrolled

36

before-after study attributed improvements to possible Hawthorne effect rather than use of

the intervention.32 Not all reported improvements were of statistical significance; others

occurred within but not between treatment groups; and in several controlled studies, observer

effect may have accounted for improvements seen in non-app users.36

This review revealed appealing app features were tracking daily healthy behaviours, self-

monitoring measurements and symptoms, appointment reminders, disease education, and

managing communication with providers. As apps become integrated into usual health care,

studies indicate that patients want more flexible options in terms of entering, editing and

viewing data. They also value customisable app features that are typically system-generated,

such as notifications and motivational messages. Further noted was the appeal of optional

interaction with other people in the game-based apps. In this review, digital products that are

targeted to self-management were viewed as increasingly routine.

Understanding and optimising enablers to uptake of self-care tools offered via mobile

platforms will continue to widen their appeal, relevance and utility. For example, within this

review, the mean age of participants was < 60 years in seven of the studies. The appeal of

mobile app use in these studies could possibly be ascribed to younger age, indicating a

selection bias. Although often assumed to be a predictor of technology appeal, age per se is

not a barrier to app use. Simple navigation that is visually clear and personalised are among

the appealing features for older app users.37 Ongoing promotion of apps to adults with CVD

must adopt such core design features. Addressing other issues such as outdated devices,

service connectivity and affordability, and tailoring cultural and linguistic adaptations where

appropriate, will continue to remove practical barriers to uptake. Notably, no studies in this

37

review reported sub-analyses based on, for example, sex or age. Further research focused on

efficacy of smartphone-based programs relative to sociodemographic variables would thus

broaden and strengthen the insights found in this review.

Several implications emerged from this review relating to the future use of mobile device-

based apps in CVD populations. In a point of difference to many apps with even nominal

interaction with health staff, an app whose design excluded this feature improved self-care

across a range of important outcomes in patients with heart failure.31 In a second example,

patients who undertook home-based cardiac rehabilitation using a smartphone app in lieu of,

not as an adjunct to, clinic-based cardiac rehabilitation, did at least as well in terms of key

outcomes of healthier diet, increased functional capacity and lowered depression scores.29 A

stand-alone app that expands the reach of evidence-based treatment, improves disease self-

management, and lowers human resource needs has great appeal. Interestingly, user

perception of human-like attributes in technology-based interventions may be what is

important. In a recent RCT of a text-message-based intervention for patients with CHD in

which message content was entirely automated, participants felt a sense of personal

connection with health staff simply due to a familiar hospital as the sender identity and

having met the research staff at recruitment, although not thereafter.38 In a review of CVD

app personalisation strategies,39 avatars that match the user’s culture and literacy level were

another appealing medium for game-based heart health programs without human

communication. Further research could explore the role of personalisation or attribution of

human qualities (known as anthropomorphism40) to stand-alone mobile apps. Cost-

effectiveness was not examined in any of the studies in this review but warrants further

research in larger studies of CVD in more diverse economies to better understand the role of

an intervention requiring fewer clinical resources.

38

The timing of app introduction relative to hospital discharge merits further investigation. App

usage as an adjunct to clinic-based cardiac rehabilitation, for example, appears to have a

reinforcing effect on adherence to secondary prevention behaviours. This suggests that app

uptake may be time-sensitive, with proximity to starting rehabilitation driving some of the

interest. Short, easily-achievable, minimally stressful tasks for patients commencing cardiac

rehabilitation may help establish and maintain ongoing commitment, compared with more

demanding challenges for those further along from hospital discharge.34 As constrains most

interventions of this nature, little is known of long-term effectiveness on hard CVD

endpoints. Clearer perhaps is that multiple behavioural domains and risk factors are

potentially modifiable, at least in the shorter term, by the dynamics of mobile-based

programs.

Attrition rates are commonly high over time for apps targeted to management of disease risk

factors41 or long term conditions, generally.42 The inconvenience of daily data entry was cited

as a factor in lower app usage and data gaps in at least one of the included studies with a

comparatively long follow-up (six months).32 App acceptability and usefulness may be

hindered by other factors, for example patient and/or clinician confidence in the reliability of

measurements obtained by app-based software. A recent study of smartphone-based

commercial heart rate monitors, for example, revealed mixed results for accuracy between

apps, particularly as heart rate increased, highlighting that perceptions of data unreliability

may undermine adoption or continuance of app use for biometric monitoring purposes.43

Another potential barrier to engagement is user trust in data security and privacy.11 In this

review, only one study33 reported on app compliance with national confidentiality laws; no

39

study evaluated patient or provider views about this issue. Reporting data security/privacy

protocols in relation to mHealth is therefore recommended.35 Integrating gamification

principles into mobile apps may increase motivation for sustaining essential but repetitive,

routine lifestyle tasks over the longer term.11, 34 In non-health gaming, priority is given to

incentives and engagement44 but core gaming principles such as a requirement for strategic

thinking, providing motivational feedback and voluntary participation could be applied for a

disease-specific context. Hence, further evaluation research may elucidate the crucial time

point(s) from the index CVD diagnosis or event at which a particular style of mobile app

most effectively impacts lifestyle decision-making and longer-term self-care behaviour.

This review has several limitations. Although every attempt was made to locate potentially

suitable studies in languages other than English, all retrieved studies for which the full text

was available were in English. The included studies are from high income countries and

therefore do not represent the diversity of settings for which CVD prevention programs

delivered via mobile apps could be beneficial. Drawbacks were identified related to the

relatively low quality of the included studies across various items within the respective

appraisal instruments used; the implications of this for the strength of evidence for app

effectiveness on many reported outcomes is acknowledged. Sample sizes were small (six of

the studies had 50 or fewer participants), compounded by significant dropouts. Only two

studies were of six months duration; five of the studies were of eight weeks duration or less.

Thus, longer term impact on risk factor control or health service use cannot be appraised and

remains a need for future research. Biases, especially selection bias, are inherent in the non-

randomised designs. Studies lacking a control group further limits association between the

identified improvements in outcomes and the app intervention. Self-report may have over-

estimated effects on health behaviour outcomes, such as smoking and physical activity. RCTs

40

of patient-directed apps are unavoidably open label and inconsistencies across reporting of

outcomes precluded statistical comparisons. Single centre studies may have further limited

representativeness. User acceptance data were in some cases reported even where studies

otherwise excluded, or were underpowered for, clinical endpoints. However, some studies

that reported clinical measures reported no user feedback; hence there may be gaps in data

about patient preferences for features of apps .App development, implementation and

maintenance cost considerations were not routinely reported, nor whether the interventions

have been adopted into routine practice. Accounts of both could be an instructive adjunct to

outcomes data. Future reviews in this dynamic area would therefore be aided by more

systematic reporting of research. A standardised reporting format, such as the sixteen-item

mobile health evidence reporting and assessment checklist35 would help overcome variations

and aid information synthesis in future reviews.

Notwithstanding these limitations, the range of outcomes for which improvements were seen

suggests a place for mobile-based apps in CVD self-management. Mobile technologies are

recognised for their potential to improve reach, efficiency and affordability.5, 11 Those who

depend on a smartphone as their only online device, for example, may stand to benefit from

otherwise inaccessible material. Apps are broadly acceptable to consumers, in part because

they employ a familiar device interface; however, by design and intention they place greater

emphasis on patient agency. Accordingly, pragmatic patient selection for such apps may

prove the more judicious approach to ensure that those most likely to benefit do so. A

randomised trial with patient preference arms for home- or clinic-based cardiac

rehabilitation45 demonstrated similar effectiveness between groups, suggesting that patient-

preferred allocation may more realistically predict success with home-based CVD self-care.

Therefore, routine process evaluation of intervention studies could also elucidate

41

characteristics of patients for whom mobile apps are likely to be suitable and successful as an

adjunct to, or a substitute for, clinic-based management.

CONCLUSIONS

Mobile-based applications with low clinician involvement are promising strategies to help

selected patients succeed with improving risk factor control and disease self-care. At service

level, they may offer a sustainable, credible, low-resource approach within the wider

technology-enabled health care environment. At the patient level, uncertainty about longer-

term program interest and impact needs to be addressed by studies of higher scientific quality

to more reliably determine whether reported improvements are app-related, and which other

factors contribute. Analyses by sociodemographic variables would improve interpretation of

outcomes. Future research should routinely encompass process evaluation data to further

refine optimal system-side features, including personal data privacy and safety; resources

needed to manage patient-generated data; and inclusion of relevant provider perspectives.

Taken together, such data inform translation and scale-up of personal technologies to enhance

CVD self-management and reduce evidence-practice gaps.

42

Author contributions: GMC, LN and JR contributed to the conception or design of the

work. GMC, LN and JM contributed to the acquisition, analysis, or interpretation of data for

the work. GMC drafted the manuscript. LN, JM and JR critically revised the manuscript. All

gave final approval and agree to be accountable for all aspects of work ensuring integrity and

accuracy.

43

REFERENCES

1. Joseph P, Leong D, McKee M, et al. Reducing the Global Burden of Cardiovascular

Disease, Part 1: The Epidemiology and Risk Factors. Circulation research. 2017; 121: 677-

94.

2. Fuster V, Kelly BB and Vedanthan R. Promoting global cardiovascular health:

moving forward. Circulation. 2011; 123: 1671-8.

3. Organisation WH. Global action plan for the preventon and control of

noncommunicable diseases 2013-2020. Geneva, Switzerland: World Health Organisation,

2013.

4. Federation. WH. Global CVD Roadmaps. 2017.

5. Riegel B, Moser DK, Buck HG, et al. Self-Care for the Prevention and Management

of Cardiovascular Disease and Stroke: A Scientific Statement for Healthcare Professionals

From the American Heart Association. Journal of the American Heart Association. 2017; 6.

6. Kamal AK, Shaikh Q, Pasha O, et al. A randomized controlled behavioral

intervention trial to improve medication adherence in adult stroke patients with prescription

tailored Short Messaging Service (SMS)-SMS4Stroke study. BMC Neurol. 2015; 15: 212.

7. Davidson TM, McGillicuddy J, Mueller M, et al. Evaluation of an mHealth

medication regimen self-management program for African American and Hispanic

uncontrolled hypertensives. Journal of Personalized Medicine. 2015; 5: 389-405.

8. Wolf A, Fors A, Ulin K, Thorn J, Swedberg K and Ekman I. An eHealth diary and

symptom-tracking tool combined with person-centered care for improving self-efficacy after

a diagnosis of acute coronary syndrome: A substudy of a randomized controlled trial. Journal

of Medical Internet Research. 2016; 18.

9. Center PR. Smartphone Ownership and Internet Usage Continues to Climb in

Emerging Economies. 2016.

44

10. Statista. Smartphones industry: Statistics & Facts. 2017.

11. Burke LE, Ma J, Azar KM, et al. Current Science on Consumer Use of Mobile Health

for Cardiovascular Disease Prevention: A Scientific Statement From the American Heart

Association. Circulation. 2015; 132: 1157-213.

12. Demiris G, Afrin LB, Speedie S, et al. Patient-centered Applications: Use of

Information Technology to Promote Disease Management and Wellness. A White Paper by

the AMIA Knowledge in Motion Working Group. Journal of the American Medical

Informatics Association : JAMIA. 2008; 15: 8-13.

13. Park LG, Beatty A, Stafford Z and Whooley MA. Mobile Phone Interventions for the

Secondary Prevention of Cardiovascular Disease. Progress In Cardiovascular Diseases.

2016; 58: 639-50.

14. Gandhi S, Chen S, Hong L, et al. Effect of Mobile Health Interventions on the

Secondary Prevention of Cardiovascular Disease: Systematic Review and Meta-analysis. The

Canadian Journal Of Cardiology. 2017; 33: 219-31.

15. Pfaeffli Dale L, Dobson R, Whittaker R and Maddison R. The effectiveness of

mobile-health behaviour change interventions for cardiovascular disease self-management: A

systematic review. European journal of preventive cardiology. 2016; 23: 801-17.

16. Beishuizen CRL, Stephan BCM, van Gool WA, et al. Web-Based Interventions

Targeting Cardiovascular Risk Factors in Middle-Aged and Older People: A Systematic

Review and Meta-Analysis. Journal Of Medical Internet Research. 2016; 18: e55-e.

17. Adler AJ, Martin N, Mariani J, et al. Mobile phone text messaging to improve

medication adherence in secondary prevention of cardiovascular disease. The Cochrane

Database Of Systematic Reviews. 2017; 4: CD011851.

18. Cajita MI, Gleason KT and Han H-R. A Systematic Review of mHealth-Based Heart

Failure Interventions. The Journal Of Cardiovascular Nursing. 2016; 31: E10-E22.

45

19. Devi R, Singh SJ, Powell J, Fulton EA, Igbinedion E and Rees K. Internet-based

interventions for the secondary prevention of coronary heart disease. The Cochrane Database

Of Systematic Reviews. 2015: CD009386.

20. Moher D, Liberati A, Tetzlaff J and Altman DG. Preferred reporting items for

systematic reviews and meta-analyses: the PRISMA statement. BMJ (Clinical research ed).

2009; 339: b2535.

21. Australia NHFo. Secondary prevention of cardiovascular disease 2010.

22. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for

assessing risk of bias in randomised trials. BMJ (Clinical research ed). 2011; 343: d5928.

23. Institute TJB. Checklist for Quasi-Experimental Studies (non-randomized

experimental studies). Adelaide, Australia: Joanna Briggs Institute, 2017.

24. Programme CAS. Qualitative Checklist. . 2017.

25. Bengtsson U, Kjellgren K, Hallberg I, Lindwall M and Taft C. Improved Blood

Pressure Control Using an Interactive Mobile Phone Support System. Journal Of Clinical

Hypertension (Greenwich, Conn). 2016; 18: 101-8.

26. Hallberg I, Ranerup A and Kjellgren K. Supporting the self-management of

hypertension: Patients' experiences of using a mobile phone-based system. Journal Of

Human Hypertension. 2016; 30: 141-6.

27. Forman DE, LaFond K, Panch T, Allsup K, Manning K and Sattelmair J. Utility and

efficacy of a smartphone application to enhance the learning and behavior goals of traditional

cardiac rehabilitation: a feasibility study. Journal Of Cardiopulmonary Rehabilitation And

Prevention. 2014; 34: 327-34.

28. Johnston N, Bodegard J, Jerström S, et al. Effects of interactive patient smartphone

support app on drug adherence and lifestyle changes in myocardial infarction patients: A

randomized study. American Heart Journal. 2016; 178: 85-94.

46

29. Varnfield M, Karunanithi M, Lee C-K, et al. Smartphone-based home care model

improved use of cardiac rehabilitation in postmyocardial infarction patients: results from a

randomised controlled trial. Heart (British Cardiac Society). 2014; 100: 1770-9.

30. Widmer RJ, Allison TG, Lerman LO and Lerman A. Digital Health Intervention as an

Adjunct to Cardiac Rehabilitation Reduces Cardiovascular Risk Factors and

Rehospitalizations. Journal Of Cardiovascular Translational Research. 2015; 8: 283-92.

31. Hägglund E, Lyngå P, Frie F, et al. Patient-centred home-based management of heart

failure. Findings from a randomised clinical trial evaluating a tablet computer for self-care,

quality of life and effects on knowledge. Scandinavian Cardiovascular Journal: SCJ. 2015;

49: 193-9.

32. Seo WK, Kang J, Jeon M, et al. Feasibility of using a mobile application for the

monitoring and management of stroke-associated risk factors. Journal Of Clinical Neurology

(Seoul, Korea). 2015; 11: 142-8.

33. Layton AM, Whitworth J, Peacock J, Bartels MN, Jellen PA and Thomashow BM.

Feasibility and Acceptability of Utilizing a Smartphone Based Application to Monitor

Outpatient Discharge Instruction Compliance in Cardiac Disease Patients around Discharge

from Hospitalization. International Journal Of Telemedicine And Applications. 2014; 2014:

415868-.

34. Dithmer M, Rasmussen JO, Grönvall E, et al. "The Heart Game": Using Gamification

as Part of a Telerehabilitation Program for Heart Patients. Games For Health Journal. 2016;

5: 27-33.

35. Agarwal S, LeFevre AE, Lee J, et al. Guidelines for reporting of health interventions

using mobile phones: mobile health (mHealth) evidence reporting and assessment (mERA)

checklist. BMJ (Clinical research ed). 2016; 352: i1174.

47

36. McCambridge J, Witton J and Elbourne DR. Systematic review of the Hawthorne

effect: New concepts are needed to study research participation effects. Journal of Clinical

Epidemiology. 2014; 67: 267-77.

37. Neubeck L, Lowres N, Benjamin EJ, Freedman SB, Coorey G and Redfern J. The

mobile revolution--using smartphone apps to prevent cardiovascular disease. Nature reviews

Cardiology. 2015; 12: 350-60.

38. Redfern J, Santo K, Coorey G, et al. Factors Influencing Engagement, Perceived

Usefulness and Behavioral Mechanisms Associated with a Text Message Support Program.

PloS one. 2016; 11: e0163929.

39. Neubeck L, Gallagher R, Farquharson B and Clark R. Apps for Cardiovascular

Disease: The Role of Avatars in Personalisation. Health Management. 2017; 17.

40. Farzanfar R. When computers should remain computers: a qualitative look at the

humanization of health care technology. Health informatics journal. 2006; 12: 239-54.

41. Flores Mateo G, Granado-Font E, Ferré-Grau C and Montaña-Carreras X. Mobile

Phone Apps to Promote Weight Loss and Increase Physical Activity: A Systematic Review

and Meta-Analysis. Journal of Medical Internet Research. 2015; 17: e253.

42. Whitehead L and Seaton P. The Effectiveness of Self-Management Mobile Phone and

Tablet Apps in Long-term Condition Management: A Systematic Review. Journal of

Medical Internet Research. 2016; 18: e97.

43. Coppetti T, Brauchlin A, Muggler S, et al. Accuracy of smartphone apps for heart rate measurement. European journal of preventive cardiology. 2017; 24: 1287-93.44. Miller AS, Cafazzo JA and Seto E. A game plan: Gamification design principles in mHealth applications for chronic disease management. Health informatics journal. 2016; 22: 184-93.45. Dalal HM, Evans PH, Campbell JL, et al. Home-based versus hospital-based rehabilitation after myocardial infarction: A randomized trial with preference arms — Cornwall Heart Attack Rehabilitation Management Study (CHARMS). International Journal of Cardiology. 2007; 119: 202-11.

48